Author: Stephen A. Boak
Date: 09:31:55 03/25/01
Go up one level in this thread
On March 25, 2001 at 10:55:02, Hans Christian Lykke wrote:
>[Event "Level=120 min. per move"]
>[Site "DK-Albertslund"]
>[Date "2001.03.16"]
>[Round "1"]
>[White "Shredder 5 At900 288MB hash"]
>[Black "Deep Fritz At900 288MB hash"]
>[Result "*"]
>[ECO "C50"]
>[PlyCount "45"]
>[EventDate "2001.??.??"]
>
1. e4 {book 0s} 1... e5 {1} 2. Bc4 {book 0s} 2... Nf6 {1} 3. d3 {book 0s} 3...
Nc6 {0} 4. Nc3 {book 0s} 4... Bc5 {0} 5. Nf3 {book 0s} 5... d6 {0} 6. Bg5 {
book 0s} 6... h6 {0} 7. Be3 {book 0s} 7... Nd4 {0} 8. Bxd4 {+0.09/17 120:00m}
8... exd4 {0} 9. Na4 {-0.01/17 120:00m} 9... Bb4+ {-0.06/18 7201} 10. c3 {
+0.02/18 120:00m} 10... dxc3 {-0.03/18 7200} 11. Bxf7+ {+0.06/17 120:00m} 11...
Kxf7 {-0.09/18 7200} 12. Qb3+ {+0.05/16 120:00m} 12... Kg6 {-0.03/17 7199} 13.
Qxb4 {-0.07/17 120:00m} 13... cxb2 {-0.06/17 7200} 14. Qxb2 {-0.02/17 120:00m}
14... Nh5 {-0.19/17 7200} 15. O-O {-0.27/16 120:00m} 15... Rf8 {-0.16/17 7202}
16. Ne1 {-0.39/16 120:00m} 16... Kh7 {-0.13/17 7200} 17. Nc3 {-0.39/17 120:00m}
17... b6 {-0.13/18 7200} 18. Qd2 {-0.39/17 120:00m} 18... Qf6 {-0.16/18 7201}
19. Nd5 {-0.32/18 120:00m} 19... Qf7 {-0.25/18 7200} 20. f3 {-0.42/17 120:00m}
20... Bb7 {-0.19/18 7200} 21. Nc3 {-0.52/17 120:00m} 21... d5 {-0.34/17 7200}
22. Nc2 {-0.64/16 120:00m} 22... c5 {-0.38/17 7200} 23. d4 {-0.66/16 120:00m} *
>
>[D]rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
>
>Shredder played the move expected by DF: 23. d4
>
>Shredder expecting 23...Nf4 24.e5
>
>DF calculating now
>
>Venlig hilsen
>Hans Christian Lykke
QUESTIONS & THOUGHTS--
Which program is producing the scores/depths/times shown above? I notice the
score keeps bouncing up and down, move by move.
From the difference in the 'times' notations, it appears to be a combination of
Deep Fritz and Shredder 5, correct?
Apparently Shredder 5 thinks Black is up approximately 2/3 of a pawn (at
present) and Deep Fritz thinks Black is only up approximately 1/3 of a pawn.
The ply depth is nearly the same (same or 1-ply different), but at least once it
has been 2-ply different.
It will be interesting to see if the score grows for Deep Fritz (how & why),
until a clear win is present.
From my own experience analyzing my human-human tournament games with Fritz6,
CM6000 & CM8000, and Rebel (Century, etc) programs, a 2/3 pawn advantage can
often be nurtured into a win (although certain types of endings are often drawn
1-pawn down such as rook endings or with bishops of opposite colors ). A 1/3
pawn advantage is much more unclear (to forecast if a win will occur).
I don't play comp-comp games, so I can not guess if, for example, a 1/3 pawn
advantage at approximately the 23rd move in a game will draw a high percentage
of the time or not.
I think in a dynamic or unbalanced position, the program that understands (you
could say calculates) the position best will have much greater chances to win.
It may find a combination the other program missed, or a way to improve the
position and pressure until a win is found, or simply win with better endgame
technique.
In the current position, Black looks better to my human eye, with:
1. a queenside pawn majority that should produce an outside (away from kings)
passed pawn.
2. a well posted (attacking) long range bishop in a position with pawns on both
sides of the board.
This is offset only a little bit by Black's knight on the edge (since it likely
will be redeployed at next move).
As for relative king safety, I think Black and White both have actual or
potential king problems. Due to Black's bishop raking into the white kingside,
I think White may have the more serious problems later on. On the other hand,
Black's king position may allow a perpetual at some future time.
Interesting position to analyze. It will be interesting to see what happens
between these two strong programs.
--Steve
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.